Institut de Recherche en Informatique de Toulouse

Systèmes d'Information Généralisés Team


Information Retrieval - Information Exploration and Visualisation Group
(SIG/EVI)


version française
Heads: B. Dousset (UPS), J. Mothe (IUFM)

The works carried out in this group aim to propose models for textual, structured, semi-structured and non structured information retrieval. Central to all these works is the notion of context. Contextual information retrieval refers to implicit or explicit knowledge about the user's intention and environment and the system itself. Our hypothese is that making certains items of the context explicit will improve the performances of IR systems. This implies defining models which characterise IR contexts, recognise a context when a user interacts whith the system and defines IR methods which are most suitable for a given context. More precisely, we integrate in our models the available resources (reference corpus, thesaurus, ontology), their format (XML, the notion of paragraph or sentence and meta-data) and the IR task performed by the user (novelty detection, granule retrieval, scientific and technological watch). To evaluate our models and systems, we participate in TREC and INEX evaluation campaigns.

Permanent Members

Taoufiq Dkaki, Assistant professor UTM
Bernard Dousset, Professor UPS (Co-Head)
Gilles Hubert, Assistant professor UPS
Josiane Mothe, Professor IUFM (Co-Head)
Maryse Salles, Assistant professor UT1
Patricia Stolf, Assistant professor IUFM

Non permanent members

Yael Champlcaux, PhD student
Ilhem Ghalamallah, PhD student
Nathalie Hernandez, Postdoc student
Gaetan Inimod, PhD student
Désiré Kompaoré, PhD student
Eloïse Loubier, PhD student
Hamza Moussa Baraou, Research Master student
Bachelin Ralalason, PhD student

Some national cooperations

Master Subjects

Links

ARIA: Association francophone de Recherche d'Information et Applications
RFIEC: Expermimental platform RFIEC

Research themes

Publications

All publications
By year

Current projects

Ws-Talk (European project): Access to information adapted to users' language

Data Mass in Astronomy (Data mass projects): Access to multi-source information in the framework on the virtual observatory (Astronomy)

ARIEL(TCAN Project) : in cooperation with ERSS, works on adapting information retrieval systems according to linguistic indicators extracted from resources (documents, queries).

Participation in evaluation programmes

TREC: Text REtrieval Conference

INEX: INitiative for the Evaluation of Xml retrieval

Completed projects

QUEST: Query reformulation for structured document retrieval
French-British research programme, with Queen Mary University of London, Mounia Lalmas, January 2002-.

eStage :
European Information Societies Technology, IST.

IRAIA : Getting Orientation in Complex Information Spaces as an Emergent Behaviour of Autonomous Information Agents
European Information Societies Technology, IST-1999-10602, March 2000-2002.

Navigation and IR on the web : new methods and new applications
French-British research programme with F. Murtagh, 1999-2001.

Tétrafusion
French-Irish research programme Forfairt/CNRS, with A. Smeaton, 1997-98.